A practical path to MCP using the APIs you already run in production
MCP Bridge is now generally available (GA), giving teams a practical way to expose existing APIs to agents without rewriting services or compromising governance. This capability is delivered as a guided experience with API Manager configuring instance data and automated policies that are then executed on MuleSoft Flex Gateway, operating at the same control plane enterprises already use to manage API traffic.
–
The agentic era is here, but for most organizations progress is slowed by a familiar reality: the APIs, and applications that powered the last wave of digital transformation were never designed with generative or agentic AI in mind.
This tension isn’t new. Every major technology shift creates the same challenge. Business-critical applications from the previous era remain essential, even as they become less suited to support what comes next.
MuleSoft customers have navigated this moment before. They’ve repeatedly turned existing IT investments into foundations for future innovation, whether through APIs, integration, or composable architectures. That experience puts them in a strong position as organizations step into the agentic era.
For most teams, the answer isn’t building or rebuilding from scratch. It’s making the API assets they already trust accessible to agents. The emerging standard for that interaction is Anthropic’s Model Context Protocol, or MCP.
So what are IT teams to do?
One option is to build MCP endpoints at the application level. This is a valid approach, and one many customers have already adopted using the MCP Connector. It provides fine-grained control over how flows, schemas, and edge cases are exposed to agents. At the same time, it requires modifying production integrations, coordinating changes across teams, retesting flows that already work, and tying agent adoption to application release cycles.
All of that effort goes into what is ultimately an interface change, and risks slowing progress rather than accelerating it. What teams really need is a way to meet agents where they are, without turning MCP adoption into another major development project.
Introducing MCP Bridge
That’s exactly why we’re excited to announce the general availability of MCP Bridge.
MCP Bridge provides an alternative path to the same outcome. Where MuleSoft’s MCP Connector offers maximum flexibility and control within individual integrations, MCP Bridge operates as an interfacing layer designed for consistency, governance, and scale across many APIs.
With MCP Bridge, teams can expose existing APIs to agents as MCP tools without rewriting services, deploying new runtimes, or weakening governance. It delivers a practical, enterprise-ready way to adopt the Model Context Protocol using the API assets already in production today!
How does MCP Bridge work?
MCP Bridge is delivered as a pre-configured set of automated policies on MuleSoft Flex Gateway via a guided user experience in API Manager, allowing organizations to enable MCP without changing existing APIs or deployment models. Rather than introducing a new runtime or modifying application code, MCP Bridge generates new MCP Server assets that are then deployed at the same gateway layer enterprises already use to manage traffic, security, and governance.
Teams define MCP tools that map to approved API operations, making those capabilities available to agents through a standardized MCP interface. When an agent invokes a tool, the request flows through Flex Gateway and is routed to the appropriate backend API methods just like any other API call. Existing authentication, authorization, rate limiting, and observability policies continue to apply, ensuring agent traffic is governed consistently with other consumers.
As a result of the MCP Bridge process a single MCP endpoint can safely span multiple backend APIs without embedding orchestration logic into agents themselves. This allows organizations to adopt MCP incrementally, exposing selected capabilities to agents while keeping full control over how and when systems participate in agent-driven workflows.
MCP Bridge is designed for enterprise environments
In practice, enterprise landscapes are rarely homogeneous. APIs are owned by different groups, operate under different constraints, and support a wide variety of consumers.
MCP Server asset generated by MCP Bridge operates at the gateway layer where enterprises already manage traffic, security, and policy. MCP tools are defined explicitly so teams remain in control of what is exposed to agents. MCP traffic flows through the same gateway as existing API traffic, policies apply consistently, and routing is handled declaratively.
A single MCP Server can expose tools that safely span multiple systems without hard-coding orchestration logic into agents themselves.
For organizations that have already adopted MCP using the connector, MCP Bridge requires no rework. Existing MCP-enabled applications continue to operate as they do today. MCP Bridge simply adds another option to produce new MCP Server assets, making it easier to standardize MCP adoption across teams while preserving the option for deeper, application-level control where needed.
Unlike approaches that treat MCP enablement as a developer-side concern, MCP Bridge aligns adoption with the responsibilities of platform, security, and API teams. Agents get a unified experience even when the systems behind them are not.
By treating MCP as an interfacing layer, organizations can experiment, learn, and expand agent use cases incrementally, picking and choosing which operations to expose agentically. Some systems may eventually become MCP-native. Others may always be exposed through APIs. Both approaches coexist under a single governance model.
MCP for the enterprise
At MuleSoft, our enterprise approach to MCP is grounded in a simple idea: AI adoption only delivers lasting value when it scales safely.
Speed alone is not enough. Exposing business-critical systems to agents without the right control and governance quickly erodes first-mover advantage. Enterprises need a way to grow deliberately, applying AI where it creates value without forcing disruption across their entire platform.
MCP Bridge is a foundational part of that strategy. Together, MCP Bridge and the MCP Connector give MuleSoft customers two compatible routes to MCP adoption: fine-tuned control at the application layer, and scalable, governed exposure at the gateway layer.
With MCP Bridge, agents gain access to reliable, well-defined capabilities. Platform teams retain control and visibility. Backend teams continue to operate systems the way they always have. Innovation moves forward without leaving stability behind.
